Hari Shanker Srivastava

Use of multiincidence angle RADARSAT-1 SAR data to incorporate the effect of surface roughness in soil moisture estimation

The proposed study offers an approach to incorporate the effect of surface roughness in the estimation of soil moisture from space without actually measuring surface roughness conditions on ground. It is required to acquire synthetic aperture radar data at low and high incidence angles, such that the soil moisture changes are negligible between the two acquisitions.

Large-Area Soil Moisture Estimation Using Multi-Incidence-Angle RADARSAT-1 SAR Data

The sensitivity of synthetic aperture radar (SAR) backscatter to soil moisture has been adequately established. However, monitoring of soil moisture over large agricultural areas is still difficult because SAR backscatter is also sensitive to other target properties like surface roughness, crop cover, and soil texture (soil type), along with its strong sensitivity to soil moisture. Hence, to develop a methodology for large-area soil moisture estimation using SAR, it is necessary to incorporate the effects of surface roughness, crop cover, and soil texture in the soil moisture retrieval model. In this paper, a methodology for soil moisture estimation over a large area is developed using a pair of low- and high-incidence-angle RADARSAT-1 SAR data over parts of Agra, Mathura, and Bharatpur districts, India, during March 1999. The methodology requires acquisition of synthetic aperture radar data at low and high incidence angles, such that the soil moisture changes are negligible between the two acquisitions. In order to demonstrate the applicability of the developed methodology, the same was validated over a different area (parts of Saharanpur and Haridwar districts, India) during March 2005. Both test sites provided the variety of agricultural heterogeneity required for development and validation of the methodology for large-area soil moisture estimation. The proposed methodology offers an approach to incorporate the effects of surface roughness, crop cover, and soil texture in the soil moisture retrieval model from the space platform, without making any assumptions on the distributions of these parameters or without knowing the actual values of these parameters on ground.

Comparative evaluation of the sensitivity of multi‐polarized multi‐frequency SAR backscatter to plant density

Interaction of synthetic aperture radar (SAR) with vegetation is volumetric in nature, hence SAR is sensitive to the variation in vegetation density. At the same time SAR is also sensitive to other target properties such as canopy structure, canopy moisture, soil moisture and surface roughness of the underlying soil. However, the sensitivity of SAR backscatter to the vegetation density depends upon the frequency, polarization and angle of incidence at which the SAR is operated. This paper provides comparative evaluation of the sensitivity of multi‐frequency and multi‐polarized SAR backscatter to the plant density of Prosopis juliflora, a thorny plant. Monitoring of P. juliflora is of importance as the state forest department introduced it to arrest the spread of desert. In carrying out this study, data from the SIR‐C/X‐SAR mission over parts of Gujarat, India, have been used. In the present study, the variation of multi‐frequency (L and C) and multi‐polarized (HH, VV and VH) SAR backscatter with plant density has been studied. The results clearly indicate that cross‐polarized SAR backscatter at longer wavelength is the appropriate choice for the quantitative retrieval of plant density.

Incorporating soil texture in soil moisture estimation from extended low‐1 beam mode RADARSAT‐1 SAR data

Sensitivity of microwaves towards soil moisture is well understood; still, development of a practical algorithm for soil moisture estimation using microwaves is difficult. This is due to the fact that along with their strong sensitivity to soil moisture, microwave signals are also sensitive to other target properties such as soil texture, surface roughness, and crop cover. In this paper, an attempt has been made to incorporate the effect of soil texture in large area soil moisture mapping using extended low‐1 beam mode RADARSAT‐1 SAR data in such a way that knowledge of soil texture is not a prerequisite.

How far sar has fulfilled its expectation for soil moisture retrieval

Microwave remote sensing is one of the most promising tools for soil moisture estimation owing to its high sensitivity to dielectric properties of the target. Many ground-based scatterometer experiments were carried out for exploring this potential. After the launch of ERS-1, expectation was generated to operationally retrieve large area soil moisture information. However, along with its strong sensitivity to soil moisture, SAR is also sensitive to other parameters like surface roughness, crop cover and soil texture. Single channel SAR was found to be inadequate to resolve the effects of these parameters. Low and high incidence angle RADARSAT-1 SAR was exploited for resolving these effects and incorporating the effects of surface roughness and crop cover in the soil moisture retrieval models. Since the moisture and roughness should remain unchanged between low and high angle SAR acquisition, the gap period between the two acquisitions should be minimum. However, for RADARSAT-1 the gap is typically of the order of 3 days. To overcome this difficulty, simultaneously acquired ENVISAT-1 ASAR HH/VV and VV/VH data was studied for operational soil moisture estimation. Cross-polarised SAR data has been exploited for its sensitivity to vegetation for crop-covered fields where as co-pol ratio has been used to incorporate surface roughness for the case of bare soil. Although there has not been any multi-frequency SAR system onboard a satellite platform, efforts have also been made to understand soil moisture sensitivity and penetration capability at different frequencies using SIR-C/X-SAR and multi-parametric Airborne SAR data. This paper describes multi-incidence angle, multi-polarised and multi-frequency SAR approaches for soil moisture retrieval over large agricultural area.

Retrieval of surface roughness using multi-polarized Envisat-1 ASAR data

Spatial distribution of surface roughness is very critical information for many application areas. Surface roughness is often characterized using statistical distribution. However, due to the huge complexity associated with spatial soil surfaces it is difficult to accurately characterize surface roughness over large areas using statistical distribution. Surface roughness influences SAR backscatter significantly and therefore for bare soil surfaces, surface roughness plays a critical role in determining the degree of depolarization of the SAR signal. In this paper surface roughness is retrieved using multi-polarized Envisat-1 ASAR data. The depolarization ratio [σ°VH − σ°VV] has been found to be very sensitive to surface roughness. This study demonstrates an approach that can be used to retrieve quantitative surface roughness values from a space platform without making any assumptions regarding distribution of surface roughness on the ground.

Performance analysis of different predictive models for crop classification across an aridic to ustic area of Indian states

Abstract The purpose of this study is to present comparative performance analysis of different machine learning algorithms for large area crop classification. Ten Indian districts with significant rabi crops viz. wheat, mustard, gram, red lentils (masoor) have been selected for the study. Most popular classical ensemble models – bagging/ARCing, random forest (RF), gradient boosting and Importance Sampled Learning Ensemble (ISLE) with traditional single model (decision tree) have been selected for comparative analysis. To incorporate dependency of large area crop in different variables viz. parent material and soil, phenology, texture, topography, soil moisture, vegetation, climate etc., 35 digital layers are prepared using different satellite data (ALOS DEM, Landsat-8, MODIS NDVI, RISAT-1, Sentinental-1A) and climatic data (precipitation, temperature). In rabi season, field survey about crop type is carried out to prepare training data. Performance is evaluated on the basis of marginal rates, F-measure and Jaccard’s coefficient of community, Classification Success Index and Agreement Coefficients. Score is calculated to rank the algorithm. RF is best performer followed by gradient boosting for crop classification. Other ensemble methods ARCing, bagging and ISLE are in decreasing order of performance. Traditional non-ensemble method decision tree scored higher than ISLE.

An effective implementation and assessment of a random forest classifier as a soil spatial predictive model

ABSTRACT Mapping the spatial distribution of soil classes is important for informing soil use and management decisions. This study aimed to effectively implement Random Forest (RF) model and to evaluate the behaviour and performance of the model for soil classification of Indian districts. Soil-forming factors, known as ‘scorpan,’ are selected as environmental covariates to tune RF model to classify 11 different soil categories. Thirty-five digital layers are prepared using different satellite data [ALOS (Advanced Land Observing Satellite) digital elevation model, Landsat-8, Moderate Resolution Imaging Spectroradiometer normalized difference vegetation index product, RISAT-1 (Radar Imaging Satellite-1), Sentinel-1A] and climatic data (precipitation and temperature) to represent scorpan environmental covariates in the study area. The RF parameters corresponding to highest Cohen’s kappa coefficient (κ) value and lowest number of random split variables are considered optimum values for RF model. Model behaviour evaluation is based on mapping accuracy, sensitivity to data set size, and noise. Two other machine-learning methods, CART (Classification and Regression Tree) decision tree (CDT) and CART ensemble bagger (CEB), are used to provide the comparative study. To access behaviour of models to the false data set, noise in training set is produced by assigning a false class to the training set in 5% increment. Comparative performance of RF model is based on quality assessment measures. To evaluate the performance of models, marginal rates, F-measure, and Jaccard’s coefficient of the community, classification success index and agreement coefficients are selected under quality assessment measures. The score is calculated to rank the algorithm. RF model shows high stability against data set reduction in comparison to other methods. The results show that the abrupt change in accuracy is only observed after 60% training data reduction in RF model; however, significant decrease in accuracy can be noted after 45% and 25% data reduction in CEB and CDT, respectively. The RF model shows comparatively the greater resistance to noise. Overall, RF model has performed better than CDT and CEB to classify soil categories in the study area. The results of this research provide new insights into the performance of RF in the context of soil class mapping.